scnc
/
quantum-espresso
mirror of https://gitlab.com/QEF/q-e.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
15,907 Commits 5 Branches 40 Tags 1.8 GiB
Commit Graph

15907 Commits

Author SHA1 Message Date
Pietro Delugas ca812c0fac [skip ci] add test for pbe noncolin magnetic case 
modified:   jobconfig
	new file:   pw_noncolin/benchmark.out.git.inp=noncolin-pbe.in
	new file:   pw_noncolin/noncolin-pbe.in
 2019-11-02 21:30:31 +01:00
giannozz 94092d5def Merge branch 'move_wannier' into 'develop' 
Two files used only in PP/ moved there from PW; makefiles updated

See merge request QEF/q-e!641
 2019-11-02 14:43:23 +00:00
Paolo Giannozzi 2d5147bacd Two files used only in PP/ moved there from PW; makefiles updated 2019-11-02 15:34:56 +01:00
giannozz 4546356c87 Merge branch 'no_wfc_rewrite' into 'develop' 
projwfc with reduced I/O

See merge request QEF/q-e!640
 2019-11-02 13:33:42 +00:00
giannozz 037af750a3 Merge branch 'fix_noncolin' into 'develop' 
Fix noncolin

See merge request QEF/q-e!639
 2019-11-02 13:33:27 +00:00
Paolo Giannozzi 117d18e240 More removal of unneeded I/O in postprocessing: wavefunctions are read directly 
from the final collected format into distributed array "evc". Everything
should work as before, including unmodified postprocessing, but no warranty.
 2019-11-02 11:10:02 +01:00
Paolo Giannozzi 01d62a61ab Script computing dependencies updated to deal with libxc module 2019-11-02 10:54:26 +01:00
Paolo Giannozzi 5f4e65c5eb projwfc: wfcs are longer read in collected format, re-written in distributed 
one. Routine read_collected_to evc deleted; routine read_file replaced by a
wrapper that has exactly the same functionalities as before. Routine projwave
in projwfc.x, using serial diagonalization, deleted: can be replaced in all
cases by the version using parallel linear algebra. Files make.depend updated.
 2019-11-01 21:53:41 +01:00
Pietro Delugas a613d9c089 Update compute_rho.f90 
IF(lsign) moved  out of the loop
 2019-11-01 11:48:20 +00:00
Pietro Delugas ea9057e7f4 adding thread parallelism in compute_rho routine 
Changes to be committed:

	modified:   PW/src/compute_rho.f90
 2019-11-01 12:09:42 +01:00
Pietro Delugas 49250a2bcf fixing noncolinear magnetic scf !! 
a temporary  if clause erroneosly commited  in the develop branch was driving
the scf crazy

 Changes to be committed:

	modified:   PW/src/gradcorr.f90
 2019-11-01 12:09:42 +01:00
giannozz f8e5811fdb Merge branch 'projwfc' into 'develop' 
Merge of parallel and serial versions almost completed

See merge request QEF/q-e!638
 2019-11-01 09:28:45 +00:00
giannozz a28e5e6b23 Merge of parallel and serial versions almost completed 2019-10-31 18:47:15 +01:00
giannozz 1895f36133 Merge branch 'test-suite' into 'develop' 
[skip-CI] test-suite again

See merge request QEF/q-e!637
 2019-10-29 17:51:18 +00:00
giannozz 5cdb7530df Dumb mistake recently added to projwfc was breaking parallel diagonalization. 
Printout also fixed.
 2019-10-29 18:42:44 +01:00
giannozz 3fd90dc1f4 [skip-CI] test-suite again 
Check on relative accuracy of pressure in test-suite removed; the threshold
for pressure convergence in one test, recently reduced, reverted to previous
value. Should be sufficient to ensure that all tests pass for qe-gpu as well.
There aren't compelling reasons to check for relative discrepancies in
pressure anyway: the absolute error should be a sufficient test.
 2019-10-29 14:47:07 +01:00
giannozz 6a54d50351 Merge branch 'fix_PHonon' into 'develop' 
DFPT+U: The headers in Hubbard-specific routines were not accurate

See merge request QEF/q-e!636
 2019-10-29 13:33:43 +00:00
Iurii Timrov b3cd57f987 DFPT+U: The headers in Hubbard-specific routines were not accurate, 
but the implementation is correct.
 2019-10-29 13:33:43 +00:00
giannozz 63daaaeec9 Merge branch 'develop' into 'develop' 
Pressure convergence threshold reduced for test vc-relax1

See merge request QEF/q-e!633
 2019-10-26 10:22:53 +00:00
giannozz 6b08dffef4 Merge branch 'patch-1' into 'develop' 
Avoiding floating point exception in some circumstances

See merge request QEF/q-e!634
 2019-10-26 10:21:48 +00:00
giannozz 54bcf0bb00 Merge branch 'develop01' into 'develop' 
Small modifications to the metal implementation in EPW

See merge request QEF/q-e!635
 2019-10-26 10:17:56 +00:00
Samuel Ponce 0664701b08 Small modifications to the metal implementation. 2019-10-26 00:01:57 +01:00
Lee Evitts 189bb42de1 When options filqf & filkf are used (and fine mesh points are subsequently disabled), variables n*f* = 0. Sometimes this can result in a floating point error when it then tries to check the modulus within the same if statement (possibly compiler dependent?). Fixed by separating out into a new if statement. 2019-10-25 16:19:56 +00:00
giannozz 80cba12598 Merge branch 'develop01' into 'develop' 
Conductivity and resistivity for metals

See merge request QEF/q-e!632
 2019-10-25 12:13:23 +00:00
Pietro Bonfa ebc3b390a8 Pressure convergence threshold reduced for test vc-relax1 2019-10-25 12:02:42 +02:00
Samuel Ponce 198915453e Merge remote-tracking branch 'gitlab/develop' into develop01 2019-10-25 10:59:38 +01:00
giannozz 5344168059 Merge branch 'update_hdf5_configure' into 'develop' 
updating configure for HDF5 usage

See merge request QEF/q-e!631
 2019-10-24 09:45:16 +00:00
Pietro Delugas 16c6c2e2f3 updating configure for HDF5 usage 
* HDF5_LIB variable in make.inc.in renamed HDF5_LIBS ( see issue #149)
* simplified syntax for HDF5 version number check in configure

 Changes to be committed:
	modified:   install/configure
	modified:   install/m4/x_ac_qe_hdf5.m4
	modified:   install/make.inc.in
 2019-10-24 11:19:27 +02:00
Samuel Ponce d576f593c8 Fix restart dos when doing metal 2019-10-23 22:46:42 +01:00
Samuel Ponce f648c869c0 Update on the metal implementation 
Add a restart test
 2019-10-23 22:23:15 +01:00
Samuel Poncé 69e327534f Merge branch 'metals' into 'develop01' 
Printing DOS when doing IBTE for metals.

See merge request sponce24/q-e!4
 2019-10-23 20:47:59 +00:00
Felix Goudreault ec40171c12 Printing DOS when doing IBTE for metals. 2019-10-23 15:54:45 -04:00
giannozz 5de5019669 Merge branch 'default_ndiag' into 'develop' 
Default for ndiag

See merge request QEF/q-e!626
 2019-10-23 18:59:31 +00:00
giannozz c9f4663d6a Merge branch 'projwfc_fix' into 'develop' 
Fix for spin-orbit case after recent changes, plus better printout

See merge request QEF/q-e!629
 2019-10-23 18:59:05 +00:00
giannozz 77fba18839 Merge branch 'cold' into 'develop' 
Upon suggestion by Nicola: more appropriate credits for cold smearing

See merge request QEF/q-e!630
 2019-10-23 18:58:45 +00:00
Paolo Giannozzi 9be701a7f5 Upon suggestion by Nicola: more appropriate credits for cold smearing 2019-10-23 20:49:34 +02:00
Paolo Giannozzi 94e6cce13c Fix for spin-orbit case after recent changes, plus better printout 2019-10-23 19:48:13 +02:00
giannozz 3e8d9e954d Merge branch 'master' into 'develop' 
A in/out fix  in tpsscx (metaGGA)

See merge request QEF/q-e!628
 2019-10-23 09:08:36 +00:00
Fabrizio Ferrari ddd8c59bcf Another in/out fix in tpsscx_spin (mGGA) 2019-10-23 08:49:12 +00:00
giannozz b45e201dee Merge branch 'dry_run' into 'develop' 
New possibility of "dry run"

See merge request QEF/q-e!627
 2019-10-22 20:32:00 +00:00
Paolo Giannozzi 85be0a3a0c New possibility of "dry run" 
If nstep=0, pw.x performs a quick "dry run" with fast initializations and
nothing else. Simpler alternative to starting with a "prefix".EXIT file
present in the scratch directory. The nstep=0 case was not usable anyway.
 2019-10-22 22:03:50 +02:00
Samuel Poncé 1d01743360 Merge branch 'metals' into 'develop01' 
Fixed bugs for GNU compilers in metal implementation of IBTE in EPW.

See merge request sponce24/q-e!3
 2019-10-22 17:39:12 +00:00
Felix Goudreault 131c312cf1 Fixed bugs for GNU compilers in metal implementation of IBTE in EPW. 
- Fixed closing file bug in IBTE.
- Cleaned up more in the printing routines of IBTE.
- Changed benchmark for the GNU compiler output for
epw_metals/epw_ibte.in test.
 2019-10-22 11:43:39 -04:00
giannozz ca4e1ab597 oops ... 2019-10-22 16:24:26 +02:00
giannozz b95638c980 Default maximum number of processor for linear algebra set to the number of 
available processors and not half that much (if using ScaLAPACK), as proposed
by Ye Luo in issue #8.
 2019-10-22 16:08:36 +02:00
giannozz 14d4f575cb Merge branch 'master' into 'develop' 
M06L mGGA functional fixed

See merge request QEF/q-e!625
 2019-10-21 13:09:52 +00:00
fabrizio22 9e373fa94b M06L mGGA functional fixed 2019-10-21 11:53:12 +02:00
giannozz 8053f4c004 Merge branch 'projwfc' into 'develop' 
projwfc.x output unified across all cases

See merge request QEF/q-e!624
 2019-10-20 08:16:44 +00:00
giannozz 78015dd746 Merge branch 'develop' into 'develop' 
Add xanes_file card to namelist to allow the user to change the default name of the xanes.dat file.

See merge request QEF/q-e!623
 2019-10-20 06:45:37 +00:00
Zack Gainsforth 8c194450df Documentation for xanes_file card. Remove from input_xspectra since it was already in plot. 2019-10-19 15:18:20 -07:00